Comments (Donald F. Klein)


The discussion between Ban and Blackwell   misses crucial current issues.  “Conflict of interest“ rose to public interest  when it became apparent that  Pharma publications were regularly more outcome positive than independent studies. This   led to the suspicion of bias but with no way to prove it, since data were sacrosanct. Therefore suspicion was diverted onto the basically problematic, ad hominem approach of authors declaring income sources. This miscarried repair diverted from the basic issue “Is there really data bias?”

This issue can only be met by independent data analysis at the patient level. If a therapeutic claim is made, shouldn’t the data supporting that claim be available for independent analysis? Otherwise, peer review is helpless since it only has data summaries and inferential statistics and implicit trust in their relevance and accuracy.

That is exactly the highly charged debate going on with regard to the initially forward looking policies of the European Medicines Agency. Their web site yields worthwhile, detailed access to the EMA positions.      

However, the move to demand public access to patient level data is now stymied in court by firms claiming that such disclosure causes economic loss. The European Ombudsman has   already declared that public health issues trump questionable economic losses. Recently, it looks like EMA is backtracking. Still ambiguous re decisions but the concerns of Pharma may prove decisive. Stay tuned.     

Ira Glick and I have also addressed these issues in our paper, Klein DF, Glick ID: Conflict of interest, journal review, and publication policy, published in  Neuropsychopharmacology. 2008 Dec; 33 (13): 3023-6.  My point is that both Ban and Blackwell could have improved their rather abstract discussions by reference to the current legal and   judicial struggle for and against open access, as well as citing the various activist groups.


Donald F. Klein

July 10, 2014